Predicting freeze-thaw damage using tipping point analysis of strain data

McAlorum, J. and Mckeeman, I. and Perry, M. and Clayburn, L. and Niewczas, P. (2017) Predicting freeze-thaw damage using tipping point analysis of strain data. In: SMAR2017, 2017-09-13 - 2017-09-15, ETH Zurich.

[thumbnail of McAlorum-etal-SMAR-2017-Predicting-freeze-thaw-damage-using-tipping-point]
Preview
 Text. Filename: McAlorum_etal_SMAR_2017_Predicting_freeze_thaw_damage_using_tipping_point.pdf
Final Published Version
Download (570kB)| Preview

Abstract

This paper demonstrates how Tipping Point Analysis can be used to predict the onset of structural strain, induced by ice formation. In civil structures, water-ice transitions present many potential issues that can lead to structural damage or plant shutdowns. Examples include freeze-thaw damage in concrete and masonry, and ice build-up on moving parts such as wind turbine blades. Early indication of ice formation could prevent irreparable damage if this information could be used to actuate de-icing procedures. The transition considered in this work is the strain induced in a polypropylene container by the volume change of water as it freezes, measured using surface-mounted fibre-optic strain sensors. This first order phase transition can be detected early on using degenerate fingerprinting, which identifies “slowing down” of the noise prior to the critical point of the transition. Water was supercooled which, at freezing, causes a rapid increase in temperature, presenting an identifiable specific transition point for reference. The analysis was able to consistently predict freezing around 5-10 minutes prior to the transition. A linear relationship was found between mass calculated from the calorimetric equation and mass from experimental measurements. Strain could not be estimated from this mass, since the random process of freezing in an open top container causes an irregular distribution of force. These tests will allow the method and the model to be continually developed towards a more practical application.

CORE (COnnecting REpositories)